softfloat.c

来自「omap3 linux 2.6 用nocc去除了冗余代码」· C语言 代码 · 共 1,878 行 · 第 1/5 页

         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) {
            float_raise( float_flag_invalid );
        }
        return 0;
    }
    return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 );

}

/*
-------------------------------------------------------------------------------
Returns 1 if the single-precision floating-point value `a' is less than or
equal to the corresponding value `b', and 0 otherwise.  The comparison is
performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic.
-------------------------------------------------------------------------------
*/
flag float32_le( float32 a, float32 b )
{
    flag aSign, bSign;

    if (    ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) )
         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        float_raise( float_flag_invalid );
        return 0;
    }
    aSign = extractFloat32Sign( a );
    bSign = extractFloat32Sign( b );
    if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 );
    return ( a == b ) || ( aSign ^ ( a < b ) );

}

/*
-------------------------------------------------------------------------------
Returns 1 if the single-precision floating-point value `a' is less than
the corresponding value `b', and 0 otherwise.  The comparison is performed
according to the IEC/IEEE Standard for Binary Floating-point Arithmetic.
-------------------------------------------------------------------------------
*/
flag float32_lt( float32 a, float32 b )
{
    flag aSign, bSign;

    if (    ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) )
         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        float_raise( float_flag_invalid );
        return 0;
    }
    aSign = extractFloat32Sign( a );
    bSign = extractFloat32Sign( b );
    if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 );
    return ( a != b ) && ( aSign ^ ( a < b ) );

}

/*
-------------------------------------------------------------------------------
Returns 1 if the single-precision floating-point value `a' is equal to the
corresponding value `b', and 0 otherwise.  The invalid exception is raised
if either operand is a NaN.  Otherwise, the comparison is performed
according to the IEC/IEEE Standard for Binary Floating-point Arithmetic.
-------------------------------------------------------------------------------
*/
flag float32_eq_signaling( float32 a, float32 b )
{

    if (    ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) )
         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        float_raise( float_flag_invalid );
        return 0;
    }
    return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 );

}

/*
-------------------------------------------------------------------------------
Returns 1 if the single-precision floating-point value `a' is less than or
equal to the corresponding value `b', and 0 otherwise.  Quiet NaNs do not
cause an exception.  Otherwise, the comparison is performed according to the
IEC/IEEE Standard for Binary Floating-point Arithmetic.
-------------------------------------------------------------------------------
*/
flag float32_le_quiet( float32 a, float32 b )
{
    flag aSign, bSign;
    //int16 aExp, bExp;

    if (    ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) )
         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        /* Do nothing, even if NaN as we're quiet */
        return 0;
    }
    aSign = extractFloat32Sign( a );
    bSign = extractFloat32Sign( b );
    if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 );
    return ( a == b ) || ( aSign ^ ( a < b ) );

}

/*
-------------------------------------------------------------------------------
Returns 1 if the single-precision floating-point value `a' is less than
the corresponding value `b', and 0 otherwise.  Quiet NaNs do not cause an
exception.  Otherwise, the comparison is performed according to the IEC/IEEE
Standard for Binary Floating-point Arithmetic.
-------------------------------------------------------------------------------
*/
flag float32_lt_quiet( float32 a, float32 b )
{
    flag aSign, bSign;

    if (    ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) )
         || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
       ) {
        /* Do nothing, even if NaN as we're quiet */
        return 0;
    }
    aSign = extractFloat32Sign( a );
    bSign = extractFloat32Sign( b );
    if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 );
    return ( a != b ) && ( aSign ^ ( a < b ) );

}

/*
-------------------------------------------------------------------------------
Returns the result of converting the double-precision floating-point value
`a' to the 32-bit two's complement integer format.  The conversion is
performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic---which means in particular that the conversion is rounded
according to the current rounding mode.  If `a' is a NaN, the largest
positive integer is returned.  Otherwise, if the conversion overflows, the
largest integer with the same sign as `a' is returned.
-------------------------------------------------------------------------------
*/
int32 float64_to_int32( struct roundingData *roundData, float64 a )
{
    flag aSign;
    int16 aExp, shiftCount;
    bits64 aSig;

    aSig = extractFloat64Frac( a );
    aExp = extractFloat64Exp( a );
    aSign = extractFloat64Sign( a );
    if ( ( aExp == 0x7FF ) && aSig ) aSign = 0;
    if ( aExp ) aSig |= LIT64( 0x0010000000000000 );
    shiftCount = 0x42C - aExp;
    if ( 0 < shiftCount ) shift64RightJamming( aSig, shiftCount, &aSig );
    return roundAndPackInt32( roundData, aSign, aSig );

}

/*
-------------------------------------------------------------------------------
Returns the result of converting the double-precision floating-point value
`a' to the 32-bit two's complement integer format.  The conversion is
performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic, except that the conversion is always rounded toward zero.  If
`a' is a NaN, the largest positive integer is returned.  Otherwise, if the
conversion overflows, the largest integer with the same sign as `a' is
returned.
-------------------------------------------------------------------------------
*/
int32 float64_to_int32_round_to_zero( float64 a )
{
    flag aSign;
    int16 aExp, shiftCount;
    bits64 aSig, savedASig;
    int32 z;

    aSig = extractFloat64Frac( a );
    aExp = extractFloat64Exp( a );
    aSign = extractFloat64Sign( a );
    shiftCount = 0x433 - aExp;
    if ( shiftCount < 21 ) {
        if ( ( aExp == 0x7FF ) && aSig ) aSign = 0;
        goto invalid;
    }
    else if ( 52 < shiftCount ) {
        if ( aExp || aSig ) float_raise( float_flag_inexact );
        return 0;
    }
    aSig |= LIT64( 0x0010000000000000 );
    savedASig = aSig;
    aSig >>= shiftCount;
    z = aSig;
    if ( aSign ) z = - z;
    if ( ( z < 0 ) ^ aSign ) {
 invalid:
        float_raise( float_flag_invalid );
        return aSign ? 0x80000000 : 0x7FFFFFFF;
    }
    if ( ( aSig<<shiftCount ) != savedASig ) {
        float_raise( float_flag_inexact );
    }
    return z;

}

/*
-------------------------------------------------------------------------------
Returns the result of converting the double-precision floating-point value
`a' to the 32-bit two's complement unsigned integer format.  The conversion
is performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic---which means in particular that the conversion is rounded
according to the current rounding mode.  If `a' is a NaN, the largest
positive integer is returned.  Otherwise, if the conversion overflows, the
largest positive integer is returned.
-------------------------------------------------------------------------------
*/
int32 float64_to_uint32( struct roundingData *roundData, float64 a )
{
    flag aSign;
    int16 aExp, shiftCount;
    bits64 aSig;

    aSig = extractFloat64Frac( a );
    aExp = extractFloat64Exp( a );
    aSign = 0; //extractFloat64Sign( a );
    //if ( ( aExp == 0x7FF ) && aSig ) aSign = 0;
    if ( aExp ) aSig |= LIT64( 0x0010000000000000 );
    shiftCount = 0x42C - aExp;
    if ( 0 < shiftCount ) shift64RightJamming( aSig, shiftCount, &aSig );
    return roundAndPackInt32( roundData, aSign, aSig );
}

/*
-------------------------------------------------------------------------------
Returns the result of converting the double-precision floating-point value
`a' to the 32-bit two's complement integer format.  The conversion is
performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic, except that the conversion is always rounded toward zero.  If
`a' is a NaN, the largest positive integer is returned.  Otherwise, if the
conversion overflows, the largest positive integer is returned.
-------------------------------------------------------------------------------
*/
int32 float64_to_uint32_round_to_zero( float64 a )
{
    flag aSign;
    int16 aExp, shiftCount;
    bits64 aSig, savedASig;
    int32 z;

    aSig = extractFloat64Frac( a );
    aExp = extractFloat64Exp( a );
    aSign = extractFloat64Sign( a );
    shiftCount = 0x433 - aExp;
    if ( shiftCount < 21 ) {
        if ( ( aExp == 0x7FF ) && aSig ) aSign = 0;
        goto invalid;
    }
    else if ( 52 < shiftCount ) {
        if ( aExp || aSig ) float_raise( float_flag_inexact );
        return 0;
    }
    aSig |= LIT64( 0x0010000000000000 );
    savedASig = aSig;
    aSig >>= shiftCount;
    z = aSig;
    if ( aSign ) z = - z;
    if ( ( z < 0 ) ^ aSign ) {
 invalid:
        float_raise( float_flag_invalid );
        return aSign ? 0x80000000 : 0x7FFFFFFF;
    }
    if ( ( aSig<<shiftCount ) != savedASig ) {
        float_raise( float_flag_inexact );
    }
    return z;
}

/*
-------------------------------------------------------------------------------
Returns the result of converting the double-precision floating-point value
`a' to the single-precision floating-point format.  The conversion is
performed according to the IEC/IEEE Standard for Binary Floating-point
Arithmetic.
-------------------------------------------------------------------------------
*/
float32 float64_to_float32( struct roundingData *roundData, float64 a )
{
    flag aSign;
    int16 aExp;
    bits64 aSig;
    bits32 zSig;

    aSig = extractFloat64Frac( a );
    aExp = extractFloat64Exp( a );
    aSign = extractFloat64Sign( a );
    if ( aExp == 0x7FF ) {
        if ( aSig ) return commonNaNToFloat32( float64ToCommonNaN( a ) );
        return packFloat32( aSign, 0xFF, 0 );
    }
    shift64RightJamming( aSig, 22, &aSig );
    zSig = aSig;
    if ( aExp || zSig ) {
        zSig |= 0x40000000;
        aExp -= 0x381;
    }
    return roundAndPackFloat32( roundData, aSign, aExp, zSig );

}


/*
-------------------------------------------------------------------------------
Rounds the double-precision floating-point value `a' to an integer, and
returns the result as a double-precision floating-point value.  The
operation is performed according to the IEC/IEEE Standard for Binary
Floating-point Arithmetic.
-------------------------------------------------------------------------------
*/
float64 float64_round_to_int( struct roundingData *roundData, float64 a )
{
    flag aSign;
    int16 aExp;
    bits64 lastBitMask, roundBitsMask;
    int8 roundingMode;
    float64 z;

    aExp = extractFloat64Exp( a );
    if ( 0x433 <= aExp ) {
        if ( ( aExp == 0x7FF ) && extractFloat64Frac( a ) ) {
            return propagateFloat64NaN( a, a );
        }
        return a;
    }
    if ( aExp <= 0x3FE ) {
        if ( (bits64) ( a<<1 ) == 0 ) return a;
        roundData->exception |= float_flag_inexact;
        aSign = extractFloat64Sign( a );
        switch ( roundData->mode ) {
         case float_round_nearest_even:
            if ( ( aExp == 0x3FE ) && extractFloat64Frac( a ) ) {
                return packFloat64( aSign, 0x3FF, 0 );
            }
            break;
         case float_round_down:
            return aSign ? LIT64( 0xBFF0000000000000 ) : 0;
         case float_round_up:
            return
            aSign ? LIT64( 0x8000000000000000 ) : LIT64( 0x3FF0000000000000 );
        }
        return packFloat64( aSign, 0, 0 );
    }
    lastBitMask = 1;
    lastBitMask <<= 0x433 - aExp;
    roundBitsMask = lastBitMask - 1;
    z = a;
    roundingMode = roundData->mode;
    if ( roundingMode == float_round_nearest_even ) {
        z += lastBitMask>>1;
        if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask;
    }
    else if ( roundingMode != float_round_to_zero ) {
        if ( extractFloat64Sign( z ) ^ ( roundingMode == float_round_up ) ) {
            z += roundBitsMask;
        }
    }
    z &= ~ roundBitsMask;
    if ( z != a ) roundData->exception |= float_flag_inexact;
    return z;

}

/*
-------------------------------------------------------------------------------
Returns the result of adding the absolute values of the double-precision